A new algorithm to identify difficult-to-replicate regions of the human genome

Our mathematical model creates a prediction of when each position of the human genome is replicated (red curve). Comparing the model's prediction with DNA replication timing data (grey curve) highlights points of deviation where replication proceeds more slowly than expected, indicating genomic regions that are particularly vulnerable to replication stress.

A manuscript entitled "DNA replication timing reveals genome-wide features of transcription and fragility" led by our postdoc Francisco Berkemeier has just been published in Nature Communications. In this work, we present a high-resolution mathematical model that captures the intricate relationship between DNA replication timing, origin firing, transcription, and chromatin organisation. The mathematical model creates a "null hypothesis" of how genome replication is expected to proceed in the absence of any stress or perturbations. Comparing the model's prediction with replication timing data reveals that genomic regions where the prediction and data diverge often coincide with fragile sites and long genes, while regions of strong concordance are linked to open chromatin and active promoters. In addition to mapping these dynamics, our approach provides a new algorithm for researchers to uncover new regions of the genome that may be susceptible to breaking and rearranging under conditions of replication stress, a hallmark of cancer.

We are very grateful to the Leverhulme Trust for supporting this work, as well as our excellent high-performance computing facility at the Cambridge Service for Data-Driven Discovery (CSD3) which made this work possible.

Team

Francisco Berkemeier

Peter Cook

Michael Boemo